The Moderate Resolution Imaging Spectroradiometer (MODIS)
C6 L3, Clouds and the Earth's Radiant Energy System (CERES) Edition-4 L3
products, and the European Centre for Medium-Range Weather Forecasts (ECMWF)
ERA-Interim reanalysis data are employed to systematically study
aerosol–cloud correlations over three anthropogenic aerosol regions and
their adjacent oceans, as well as explore the effect of retrieval artifacts
and underlying physical mechanisms. This study is confined to warm phase and
single-layer clouds without precipitation during the summertime (June, July,
and August). Our analysis suggests that cloud effective radius (CER) is
positively correlated with aerosol optical depth (AOD) over land (positive
slopes), but negatively correlated with aerosol index (AI) over oceans
(negative slopes) even with small ranges of liquid water path
(quasi-constant). The changes in albedo at the top of the atmosphere (TOA)
corresponding to aerosol-induced changes in CER also lend credence to the
authenticity of this opposite aerosol–cloud correlation between land and
ocean. It is noted that potential artifacts, such as the retrieval biases of
both cloud (partially cloudy and 3-D-shaped clouds) and aerosol, can result
in a serious overestimation of the slope of CER–AOD/AI. Our results show
that collision–coalescence seems not to be the dominant cause for positive
slope over land, but the increased CER caused by increased aerosol might
further increase CER by initializing collision–coalescence, generating a
positive feedback. By stratifying data according to the lower tropospheric
stability and relative humidity near cloud top, it is found that the
positive correlations more likely occur in the case of drier cloud top and
stronger turbulence in clouds, while negative correlations occur in the case of
moister cloud top and weaker turbulence in clouds, which implies entrainment
mixing might be a possible physical interpretation for such a positive
CER–AOD slope.
Introduction
Atmospheric particles influence the Earth's radiation budget and hence
climate change directly by scattering and absorbing solar radiation and
indirectly by acting as cloud condensation nuclei (CCN), altering cloud
properties and precipitation (Twomey, 1977; Albrecht, 1989). The latter
historically has been referred to as the aerosol indirect effect, and more
recently as the effective radiative forcing due to aerosol–cloud
interactions, which continue to have the largest uncertainty in assessing the
anthropogenic contribution to present climate change (IPCC, 2013). An
increase in CCN number concentration will generate a cloud that consists of
more but smaller drops – under constant cloud liquid water path (LWP). The
consequence is scattering of more solar radiation back to space. The
decrease in cloud effective radius (CER) with increasing aerosol
concentrations was historically termed aerosol first indirect effect (AIE),
cloud albedo effect, or Twomey effect (Twomey, 1977: Feingold et al., 2003).
There have been many observational evidences from in situ aircraft
measurements (Pawlowska and Brenguier, 2000; Wilcox et al., 2006; Roberts
et al., 2008; Kleinman et al., 2012; Werner et al., 2014) and ground-based (Feingold
et al., 2003; Kim et al., 2003; Garrett et al., 2004; Qiu et al., 2017) and
satellite-based (Nakajima et al., 2001; Bréon et al., 2002; Kaufman et
al., 2005; Koren et al., 2005, 2010; Quaas et al., 2008; Costantino and
Bréon, 2010, 2013; Lihavainen et al., 2010; Chen et al., 2014;
Christensen et al., 2016) remote sensing in support of this negative
correlation between aerosol concentrations and CER. Moreover, solid
evidence for the Twomey effect was also found from aerosol-induced changes in
top-of-atmosphere (TOA) albedo. Su et al. (2010) found a significant
increase in TOA albedo associated with continental aerosols relative to
those associated with oceanic aerosols under all LWP ranges when cloud
fraction is constrained; similar results have been reported by Chen et al. (2014) and Christensen et al. (2016). However, relationships of aerosol
optical depth and aerosol index with cloud albedo, cloud fraction, and cloud
liquid water path have been shown to be difficult to interpret due to the
confounding influence of relative humidity fluctuations that impact both
quantities (Quaas et al., 2009, 2010; Gryspeerdt et al., 2014, 2016).
In addition to the explorations on aerosol–cloud interactions at the cloud
top (from satellite-based remote-sensing instruments) and cloud base (from
ground-based remote-sensing instruments), the role of vertical observations
in the cloud–aerosol–precipitation interaction has been studied by using
satellite-based radar/lidar. For example, the measurements provided by
CALIOP sensors were widely used to obtain the respective position of aerosol
and cloud layers (Costantino and Bréon, 2010, 2013; Zuidema, et al.,
2016; Liu et al., 2017), which can improve the estimation of the amount of
aerosols that actually enter the cloud. The CPR (cloud profiling radar)
on board CloudSat is able to penetrate optically thick clouds layers (Wang et
al., 2013), and thus has been used to differentiate cloud regimes (Peng et
al., 2016; Christensen et al., 2016) and investigate the response of cloud
vertical structure to aerosols (Chen et al., 2016). Additionally, the
aerosol-induced changes in the vertical structure of precipitation were also
explored by using the Tropical Rainfall Measuring Mission (TRMM) products
(Wall et al., 2014; Li et al., 2016; Guo et al., 2018).
In addition to the widely observed Twomey effect, positive correlations
between aerosol concentrations and CER were also found in some regions, such
as southeastern US and southeastern China (Yuan et al., 2008), eastern China
(Tang et al., 2014; Wang et al., 2014, 2015; Liu et al., 2017), and India
(Panicker et al., 2010; Manoj et al., 2012). Overall, positive correlations
occur over land while negative correlations dominate over ocean (Grandey and
Stier, 2010; Ma et al., 2018). The lack of consensus on these relationships
motivates further exploration of underlying physical reasons for these
opposite correlations.
It is acknowledged that quantification of aerosol first indirect effect is
highly uncertain due to various influencing factors, including (1) aerosol
microphysics such as size (Feingold et al., 2001; Dusek et al., 2006; Zhang
et al., 2011), chemical composition (Nenes et al., 2002; Lance et al., 2004;
Ervens et al., 2005; McFiggans et al., 2006; Almeida et al., 2014), and
mixing state (Wang et al., 2008, 2010), (2) meteorological conditions such
as vertical velocity (Koren et al., 2010; Lu et al., 2012), lower
tropospheric stability (Wang et al., 2014; Saponaro et al., 2017; Ma et al.,
2018), wind shear (Fan et al., 2009), and precipitable water vapor (Yuan et
al., 2008; Qiu et al., 2017), (3) cloud types (Gryspeerdt and Stier, 2012;
Chen et al., 2016), and (4) vertical overlapping status of aerosol and cloud
layers (Costantino and Bréon, 2010, 2013; Huang et al., 2015; Liu et
al., 2017). It is extremely difficult to completely isolate the response of
CER to aerosol perturbations from the abovementioned influencing factors.
Yuan et al. (2008) examined the positive correlation between CER and aerosol
optical depth (AOD) by using Moderate Resolution Imaging Spectroradiometer
(MODIS) satellite products, and speculated that slightly soluble organics
particles, which induced a decrease in aerosol activation, might be a
possible explanation for the positive correlation. Wang et al. (2014)
explored AIE over eastern China by using both MODIS satellite and National
Centers for Environmental Prediction (NCEP) reanalysis data and found that
positive correlations are more likely to occur under unstable atmosphere
conditions. Ma et al. (2018) employed MODIS satellite products and
ERA-Interim reanalysis data to systematically explore the impact of
meteorological conditions on the correlations over the major industrial
regions and over relatively clean oceans, and concluded that positive
correlations are more likely to correspond to relatively high cloud-top
height (CTH) and low lower tropospheric stability (LTS), while negative
relationships were predominantly found for low CTH and high LTS. Tang et al. (2014) pointed out that covariation in wind field and relative humidity in
the North China Plain may contribute to such positive correlations; that is,
relatively wet and polluted southerly wind leads to simultaneous increases in
both AOD and CER, while dry and clean northerly wind results in coincident
decreases in AOD and CER. Gryspeerdt and Stier (2012) reported that the
strongest positive correlation occurs in the shallow cumulus cloud regimes,
while the negative one is analyzed for the stratiform cloud regimes. By
using the profiles from the CALIPSO lidar that is able to identify the
respective position of aerosol and cloud layers, Costantino and Bréon (2010, 2013) found that the aerosol indirect effect is stronger for
well-mixed aerosol and cloud layers than for separated ones.
In addition to the real physical–chemical processes stated above, the positive
correlations between AOD (or aerosol index, AI) and CER may also result from
artificial correlations due to retrieval biases. The MODIS aerosol
retrieval algorithm is conducted only for clear pixels determined by a cloud
mask. AOD could be overestimated due to either cloud contaminations where
spurious clouds might be present in pixels that are erroneously identified
as completely clear pixels (Kaufman et al., 2005; Remer et al., 2005; Zhang
et al., 2005) or cloud adjacency (or 3-D) effect where cloud-free pixels
are brightened by reflected light from surrounding clouds (Cahalan et al.,
2001; Wen et al., 2006, 2007; Várnai and Marshak, 2009). Moreover, cloud
retrievals applied to partially cloudy and 3-D-shaped cloud pixels are
expected to deviate from the retrieval assumptions of overcast homogenous
cloud and 1-D plane-parallel radiative transfer, and tend to result in an
overestimation of CER (Han et al., 1994; Coakley et al., 2005; Matheson et
al., 2006; Zhang and Platnick, 2011; Zhang et al., 2012; Grosvenor et al.,
2018). Therefore, covariation in biases in CER and AOD (AI) may incur a
false correlation between the two variables.
As stated above, aerosol–cloud correlations derived from satellite remote
sensing are potentially veiled by large variations in both retrieval biases
and meteorological conditions. By employing MODIS and CERES satellite data
as well as ERA-Interim reanalysis data, this study aims to (a) examine
whether the positive relationship between AOD and CER over land is true, (b) assess how and to what extent the satellite retrieval biases may affect the
satellite-diagnosed aerosol–cloud correlations, and (c) explore the
underlying physical mechanisms. This paper is organized as follows: the data
descriptions of both satellite and reanalysis and data processing are
presented in Sect. 2, and major findings are discussed in Sect. 3. A summary
and discussion is given in Sect. 4.
Data
In this study, three regions with strong anthropogenic emissions over land,
namely, eastern China (EC), eastern US (EU), and western Europe (WE), as well as
their neighboring oceans (ECO, EUO, and WEO, respectively), are chosen to
systematically examine aerosol–cloud correlations under different
anthropogenic emissions and dynamic and thermodynamic conditions (Fig. 1).
The data used include aerosol and cloud properties gathered from the
MODIS Aqua Collection 6 Level-3 (L3) daily product (Levy et al., 2013;
Platnick et al., 2017), albedo at the top of the atmosphere (TOA) obtained
from the Clouds and the Earth's Radiant Energy System (CERES) Aqua Edition-4
L3 product (Wielicki et al., 1996), as well as meteorological variables
extracted from the European Centre for Medium-Range Weather Forecasts
(ECMWF) ERA-Interim reanalysis data (Dee et al., 2011). The 14 consecutive
years (2003–2016) of daily 1∘×1∘ gridded
data are used for statistical analysis in order to obtain statistically
significant results from enough samples.
Regions analyzed in this study: three anthropogenic
regions (EC, EU, and WE) and their adjacent oceans (ECO, EUO, and WEO).
Satellite data
The MODIS C6 L3 product provides AOD retrieved at several wavelengths
globally (Levy et al., 2013; Sayer et al., 2013), which has been validated
extensively (Remer et al., 2005; Tripathi et al., 2005). In this study, the
AOD products retrieved from the Dark Target (DT) algorithm are used. Aerosol
index (AI), in comparison to AOD, is believed to be a better proxy for CCN
since it contains the information of aerosol size (Nakajima et al., 2001;
Stier, 2016). The AI can be derived on the basis of AOD and Ångström
exponent (AE), with the former provided directly by the MODIS product and
the latter calculated from AOD at wavelengths of 460 and 660 nm. However, the
MODIS retrieval of AE over land may be problematic (Levy et al., 2013; Sayer
et al., 2013). For Collection 6 products, several previous studies have
found little quantitative skill in MODIS-retrieved aerosol size parameters
over land (Levy et al., 2010; Mielonen et al., 2011). For this reason, we
use AOD and AI as proxies for CCN over land and ocean, respectively.
The MODIS C6 L3 product provides cloud macrophysical parameters, including
cloud fraction (CF), cloud-top temperature, cloud-top pressure, and cloud-top height, as well as cloud microphysical parameters (CER, LWP, and cloud
optical depth) with statistics (mean, minimum, maximum, and standard
deviation) at three wavelengths (1.6, 2.1, and 3.7 µm) for individual
cloud phases (liquid, ice, and undetermined) separately. We filtered the
MODIS cloud data according to the criteria employed by Saponaro et al. (2017) to ensure the data used for analysis are only limited to warm liquid
phase and single-layer clouds, and non-precipitating cases. In previous
versions, before MODIS C6, all pixels identified as partly cloudy (either
partially cloud-covered or at cloud edge) were restored to clear sky, and
the corresponding cloud retrievals were missing; that is, cloud retrievals
were only performed in overcast cloudy pixels. In C6, however, the
retrievals of cloud microphysical properties are now attempted on these
pixels, and successful retrievals are reported in the Level-2 product and
aggregated to the Level-3 product, which are reported separately in partly
cloudy (PCL) science data sets (SDSs) and are segregated from the normal
“overcast” SDSs. Therefore, the simultaneous availability of PCL and
normal overcast SDSs provides a great opportunity to investigate the
effect of partly cloudy retrievals on aerosol–cloud correlations.
The CERES SSF1deg Edition-4 product provides radiative fluxes and albedo at
TOA for all-sky and clear conditions in the longwave, shortwave, and window
regions (Loeb et al., 2005). The albedo at TOA (plenary albedo) is the ratio
of broadband (0.2–5 µm) shortwave radiation reflected and the incoming solar flux
at the top of the atmosphere. In comparison with the previous version, the
CERES instrument calibration and CERES instrument spectral response function
corrections have been improved in Edition-4 (Doelling et al., 2016a, b).
The satellite-based measurements can provide long-term data sets of aerosols
and clouds in high spatial and temporal resolution on a global scale, and
have been believed to be a promising tool for the study of aerosol–cloud
interactions. However, several inherent limitations prevent a definite
quantitative interpretation of aerosol indirect effects. The relationship
between aerosol and cloud droplet number, Nd, at cloud base
is fundamentally controlled by both the critical supersaturation of aerosol
particles, which is described by Köhler theory, and the maximum
supersaturation at cloud base, which is determined by the updraft speed and
the competition of the existing CCN for the available humidity. Nd at
cloud base is thus a function of cloud-scale, cloud-base vertical wind,
aerosol size distribution, and solubility. Nd at cloud base thus
necessarily is positively correlated to CCN, with an approximately
logarithmic relationship. In this study, we use the satellite-retrieved
cloud-top (at approximately 1–5 optical depth into the cloud,
depending on the wavelength employed to retrieve it; Platnick, 2000) CER
stratified by classes of LWP as cloud quantity. It deviates from cloud-base
Nd since
the stratification by LWP bins might not be sufficient to disentangle LWP and Nd impacts on CER;
cloud-top Nd might be related only loosely to cloud-base Nd in
case the clouds are non-adiabatic (entrain environmental air from the sides
or top), and/or coagulation of cloud droplets occurs;
the retrieved CER might be contaminated by various retrieval problems and thus is possibly only loosely related to the real cloud-top CER.
As an aerosol quantity, we use AOD/AI retrieved in the pixels determined as
cloud-free in the retrieval algorithm. This may be only loosely related to
the cloud-base CCN. Stier (2016) stated that 52 % of the area of the
globe shows correlation coefficients between CCN0.2% at cloud base
and AI below 0.5. This is largely due to
AOD/AI being a vertical integral and so might be dominated by aerosol layers
that do not affect cloud-base CCN concentrations (Stier, 2016);
the clear-sky column retrievals possibly not being representative of the
cloud-base CCN in the neighboring cloudy pixels;
AOD/AI might be loosely related to cloud-base CCN if affected by
deliquescence in high-relative-humidity environments, or if dust or other
insoluble aerosol has a substantial share of the total aerosol.
Reanalysis data
To explore the extent to which meteorological conditions affect the CER
correlation with AOD/AI, ERA-Interim reanalysis data was employed to derive
the lower tropospheric stability (LTS) and relative humidity near cloud top
(RHCT). LTS is computed as the difference between the potential
temperature at 700 hPa and at the surface (Klein and Hartmann, 1993),
representing the magnitude of the inversion strength in the lower
troposphere. Large LTS is associated with stable conditions in which
vertical mixing is suppressed. RHCT is the relative humidity over the
pressure levels closest to the cloud-top pressure. The daily 1∘×1∘ gridded reanalysis data at 14:00 local solar time
are used to match the cloud parameters obtained from MODIS L3 data.
ResultsCorrelations between CER and AOD/AI
According to Twomey (1977), the presence of anthropogenic aerosols increases
the cloud droplet number concentration but decreases CER for a constant LWP.
Since CER is a function of both AOD/AI and LWP, under conditions where LWP
changes with aerosol concentration, the variation in LWP could act to
modulate the relationship between AOD/AI and CER. Therefore, the constant
LWP assumption should be highlighted in assessing aerosol first indirect
effect. However, many previous studies did not constrain the LWP (Bréon et
al., 2002; Tang et al., 2014; Wang et al., 2014, 2015; Liu et al., 2017), or
constrained the LWP into coarse intervals (Nakajima et al., 2001; Sekiguchi
et al., 2003; Yuan et al., 2008), which would induce uncertainty in
correlating CER and AOD/AI. In our previous study (Ma et al., 2018), the CER
and AOD/AI are grouped over LWP bins with an interval of 40 g m-2, and
then a linear regression analysis with the logarithms of the CER and AOD/AI
in each LWP bin is performed (as shown in Fig. 2a). It is found that CER is
positively correlated with AOD over land (positive slopes), but negatively
correlated with AI over their adjacent oceans (negative slopes), and the
positive slopes over land become weaker while negative slopes over ocean
change slightly as LWP increases. To examine whether such intervals of LWP
bin can result in uncertainty in quantifying CER–AOD/AI correlations,
especially the positive ones over land shown in Fig. 2a, we conducted a
similar statistical analysis for the LWP versus AOD/AI within each LWP bin.
As shown in Fig. 2b, in the smallest LWP bin (20–60 g m-2, thin cloud),
LWP is more sensitive to AOD/AI, with a significant positive correlation
over land but negative correlation over ocean, which will amplify the
positive CER–AOD correlation over land and the negative CER–AI correlation
over ocean shown in Fig. 2a. For larger LWP bins (>100 g m-2), LWP is insensitive to both AOD and AI, indicating that the
increase in cloud water is governed by meteorological conditions instead of
aerosols in the case of thick clouds. Therefore, coarse intervals of the LWP bin
might be a possible cause for the appearance that the positive slopes of CER
versus AI over land become weaker as LWP increases (Ma et al., 2018).
The computed slopes of (a) CER versus aerosol loading (AOD for
land, AI for ocean) (Ma et al., 2018) and (b) LWP versus aerosol loading on a
log–log scale over six regions, in which data are stratified according to
LWP.
Joint histograms between LWP (x axis) and CER (y axis) for EC, EU,
WE, ECO, EUO, and WEO. The first and second columns show the LWP–CER joint
histograms for the low and high aerosol loading (AOD for land, AI for ocean)
cases, respectively. The histograms are normalized so each column sums to 1,
such that the histograms show the probability of observing a specific CER,
given a certain LWP. The black line indicates the mean CER at each LWP. The
third column shows the difference between the polluted and clean cases.
It is noted that the positive CER–AOD correlation over land can still be
found even within bins where LWP is insensitive to AOD, implying that the
variation in LWP within the bin may not be the dominant cause for the observed
positive relationship over land for large LWPs. However, for the smaller LWP
bin where LWP is susceptible to AOD/AI, it is still not clear whether the
sign of CER–AOD/AI correlation is interfered by LWP changes. To this end, we
constrain the LWP to a smaller range (5 g m-2) to ensure more constant
LWP. As data samples in such small intervals are insufficient to calculate
the slope of CER versus AOD/AI, the difference of CER between polluted and
clean conditions was thus employed as an alternative to the slopes. In this
study, two-dimensional (“joint”) histograms of CER and LWP are used to
show the probability distribution function of CER at each LWP bin.
Meanwhile, the 25th and 75th percentiles of the AOD/AI are computed for each
bin with a certain CER and LWP, and then the samples with AOD/AI lower than the
25th percentile and with AOD/AI greater than the 75th percentile are
classified as “clean case” and “polluted case”, respectively. The difference
between the joint histograms in polluted and clean cases demonstrates how
CER varies with AOD over land and AI over ocean in a quasi-constant LWP bin.
Figure 3 indicates that CER increases significantly with increasing LWP over
all regions, and the dependence of CER on LWP in the polluted case is weaker
than that in the clean case. By looking at the difference between the joint
histograms in polluted and clean cases (the third column in Fig. 3), it is
clearly shown that, over land, the samples of the polluted case tend to
concentrate in the larger CER bin than the clean case, while the opposite is
found over ocean. That is, as AOD/AI increases, CER increases over land but
decreases over ocean even with the tighter constraint on LWP. It can thus be
concluded that the variation in LWP within a bin does not change the sign of
CER–AOD/AI slope, although it would overestimate the correlation between CER
and AOD/AI to some extent.
Additionally, previous studies have noted a nonlinear effect of aerosols on
cloud microphysics (Bréon et al., 2002; Lohmann and Lesins, 2002; Jin and Shepherd, 2008; Gryspeerdt et al., 2017). Using the Polarization and
Directionality of the Earth's Reflectances (POLDER) satellite data, Bréon et
al. (2002) found that the slope of CER versus AI on a global scale is highly
significant for AI less than 0.15, whereas a saturation effect occurs as the
aerosol keeps rising. A similar saturation effect was also reproduced by
the simulations in global climate models (Lohmann and Lesins, 2002;
Gryspeerdt et al., 2017). In this study, such a saturation effect was found in
three ocean regions (ECO, EUO, and WEO) adjacent to the anthropogenic
regions (Fig. S1 in the Supplement), but with a higher threshold (AI = 0.3) than that
derived from the analysis on global oceans (Bréon et al., 2002). However, it
is noted that no saturation effect was found in the regions with strong
anthropogenic emissions over land (EC, EU, and WE); i.e., CER shows a
continuous increase as AOD increases (Fig. S1), suggesting that the
saturation effects do not appear to interfere with the observed positive
slopes over land.
Evidence for the positive correlation between CER and AOD over land
Although the positive correlation between CER and AOD/AI over land has been
reported by several satellite-based studies (Tang et al., 2014; Wang et al.,
2014, 2015; Liu et al., 2017; Ma et al., 2018), its reliability is still
controversial due to the limitations of satellite retrievals. The main
concerns include the following: (a) satellite-retrieved aerosol and cloud properties
continue to have biases; (b) the cloud properties observed from satellites are
confined to cloud tops rather than the entire cloud; and (c) the AI or AOD may
not actually represent the magnitude of aerosols entering the clouds.
Therefore, it is necessary to confirm whether the positive correlation is
real before discussing possible physical causes.
The difference of TOA albedo for the polluted (AOD/AI >75th percentile) and clean cases (AOD/AI <25th percentile) (Δα) over (a) EC, (b) EU, (c) WE, (d) ECO, (e) EUO, and (f) WEO, in
which data are stratified according to both LWP (x axis) and CF (y axis).
Under cloudy sky, the response of TOA in-cloud albedo at a given LWP to
aerosol changes is considered to be a more direct indicator for the Twomey
effect than CER (Su et al., 2010), given that the albedo is calculated from
CERES TOA upwelling fluxes while CER is retrieved from MODIS reflectance
based on multiple assumptions and thus tends to have larger retrieval biases
(Matheson et al., 2006; Vant-Hull et al., 2007; Zhang and Platnick, 2011;
Zhang et al., 2012). In addition, aerosol and cloud retrievals can also
interfere with each other, and thus have the potential to result in an artifact
correlation. These retrieval biases will be discussed in more detail in
Sect. 3.3.1. Therefore, using TOA albedo (α) as a proxy to examine
Twomey effect can better exclude the influence of artifacts in the
retrievals. In general, the variations in all-sky TOA albedo are controlled
by the following five parts: (a) cloud fraction (CF, representing the
horizontal extent of cloud); (b) cloud liquid water path (representing the
vertical development of a cloud); (c) cloud droplet size (that is, the
Twomey effect when considering aerosol perturbations at a certain LWP); (d) clear-sky aerosol loading and above-cloud absorbing aerosol, and (e) surface
albedo. In the presence of clouds, the aerosol-induced TOA albedo changes
are minimal compared to cloud-induced changes, and therefore (d) are not
considered in our analysis. Additionally, for a specific season and region,
surface albedo is also assumed to be uniform (Robock, 1980; Wang et
al., 2004), so the effect of (e) is ruled out. In this study, CERES TOA
albedo and AOD/AI are stratified according to CF (with an interval of 0.04)
and LWP (with an interval of 10 g m-2) in order to isolate the Twomey
effect from the interference by them, and clean case and polluted case
are distinguished according to the method used in Sect. 3.1. Thereby, for
a specific bin determined by both CF and LWP, the difference of TOA albedo
between polluted and clean conditions (Δα) can be used to
represent the changes in TOA albedo caused only by the Twomey effect (Fig. 4). The number of samples for each LWP and CF bin is shown in Fig. S2. A
positive Δα means that the observed TOA albedo has a larger
value in the polluted case than in the clean one, and the reverse is true when
Δα has a negative value. Figure 4 shows that Δα is normally negative over land (top), but positive over ocean
(bottom). The ratio of positive to total samples over EC , EU, and WE is
only 37 %, 11 %, and 13 %, respectively, which is evidently less
than that over ECO (85 %), EUO (61 %), and WEO (84 %). This result
implies that the reflected solar shortwave radiation at TOA generally tends
to decrease with AOD over land and increase with AI over ocean. As
mentioned above, under constant CF and LWP, the response of TOA albedo
to elevated aerosol loading can be interpreted as the result of changes in
CER. That is, under polluted conditions, larger droplets over land scatter
less solar radiation back to space, inducing lower TOA albedo, whereas
smaller droplets over ocean do the opposite. This finding hence lends
credibility to the positive CER–AOD correlations over land and negative
CER–AI correlations over ocean, as derived in Sect. 3.1.
Factors influencing the slope of CER versus AOD/AIArtifact correlations
Although the positive correlations between CER and AOD over land are believed
to be a real relationship as discussed in Sect. 3.2, the magnitude of this
slope may still be subject to artificial correlations due to the biases of
both aerosol retrievals (cloud contamination and cloud adjacency effect) and
cloud retrievals. In other words, if biases in CER are correlated with
biases in AOD, a false correlation may occur between the two variables.
Cloud effective radius (CER) varies with aerosol loading
(AOD for land, AI for ocean) under three liquid water path (LWP) categories
(20–60, 60–100, 100–140 g m-2). The figures from top to bottom show the
CER–AOD/AI correlation over different regions (EC, EU, WE, ECO, EUO, WEO).
The open circles and dots show the mean CER from PCL (overcast) retrieval at each
AOD (AI) bin in the case of CF >60 % (red) and CF <60 %
(blue), respectively. The slopes on a log–log scale along with the total
number of samples (bracketed) for these four categories
are also provided in each panel, in which the numbers in italic (bold)
represent PCL (overcast) retrieval.
The differences between the slope of CER versus aerosol
loading (AOD for land, AI for ocean) in PCL and in overcast retrievals in
the case of CF >60 % (ΔS|>60) and CF <60 % (ΔS|<60) under three LWP bins (g m-2).
The MODIS aerosol retrieval algorithm is applied only for clear pixels
determined by cloud mask (Remer et al., 2005). To screen out cloudy pixels,
a 3×3 standard deviation test was used to detect low clouds
(Martins et al., 2002) and IR channels to identify high clouds in addition
to the standard cloud mask product. The 20 % darkest and 50 % brightest
pixels were also removed for possible cloud contaminations. However, some pixels associated with cloud contaminations might still be erroneously
identified as clear (Kaufman et al., 2005; Remer et al., 2005; Zhang et al.,
2005), inducing an overestimation of AOD. Meanwhile, cloud-free pixels may
be brightened by reflected light from surrounding clouds, i.e., cloud
adjacency effect (Cahalan et al., 2001; Wen et al., 2006, 2007; Varnai and
Marshak, 2009), which could also overestimate AOD. The overestimation of AOD
caused by both cloud contaminations and cloud adjacency effect was found to
increase as cloud fraction (CF) increases (Zhang et al., 2005).
Additionally, cloud retrievals applied to the partly cloudy (PCL) pixels are
expected to deviate from the retrieval assumptions of overcast homogenous
cloud, and tend to overestimate CER (Han et al., 1994; Coakley et al., 2005;
Matheson et al., 2006). Therefore, a spurious positive correlation could
occur in the case of co-existence of overestimation of AOD and CER. To look into
the contribution of retrieval biases on artifact correlations, the
correlations of CER and AOD/AI derived from PCL retrievals and normal
overcast retrievals are compared, where data are stratified according to
CF in order to account for the impact of biases in aerosol retrievals (Fig. 5). Considering the limited data samples for the LWP bins of 140–180 and
180–220 g m-2, the associated statistical results are not included in
our analysis. As seen in Fig. 5, the CER–AOD/AI slopes in PCL retrieval
(circles) are significantly larger than those in overcast retrieval (dots).
It is interesting to see that, over ocean, where the Twomey effect dominates, the slope even becomes positive in the PCL case, which suggests
that CER biases in PCL retrievals can result in a serious overestimation of
the positive slope. In addition to PCL retrievals, the biases in aerosol
retrievals could also affect this positive slope. The differences between
the slope in PCL and in overcast retrievals in the case of CF >60 % (ΔS|>60) and CF <60 % (ΔS|<60) over three anthropogenic regions and their adjacent oceans
are summarized in Table 1, which clearly shows that ΔS|>60
is overall larger than ΔS|<60 under all LWP bins, implying
that the overestimation of AOD due to retrieval biases under larger CF can
amplify the positive slope caused by PCL retrievals. Therefore, in the
actual retrieval process, if the MODIS algorithm erroneously identifies PCL
pixels as overcast pixels, the artifact positive correlation will be
introduced, especially for conditions of larger CF.
In addition to PCL retrievals, the cloud retrievals for 3-D-shaped clouds
are also problematic due to deviating from the retrieval assumptions of
plane-parallel clouds (Nakajima and King, 1990). The variations in the
sun–satellite scattering geometries can interfere with satellite-measured
signals and hence induce biases in retrievals for 3-D-shaped clouds, which
are interpreted as shadowing and illumination effects (Vant-Hull et al.,
2007). Such retrieval biases in CER were found to be larger in the 2.1 µm
channel than in 3.7 µm channel (Zhang and Platnick, 2011; Zhang et al.,
2012). Our previous study (Ma et al., 2018) stated that the results of
positive slope over land and negative slope over ocean changed little when
using CER in 3.7 µm (CER3.7µm) rather than CER in 2.1 µm
(CER2.1µm) for the regression analysis. In order to examine the
effect of biases in 3-D-shaped cloud retrievals on artifact correlations,
the differences between CER2.1µm–AOD slope and CER3.7µm–AOD
slope over land and the differences between CER2.1µm–AI slope and
CER3.7µm–AI slope over ocean under all LWP bins are calculated (Fig. 6). The number of samples for each bin for regression
analysis is shown in Fig. S3. It is clearly demonstrated that the
differences are normally positive values over both land and ocean, with the
maximum difference of 0.08, which suggests that retrieval biases for 3-D-shaped cloud tend to result in a higher slope than its physically correct
value.
The differences between CER2.1µm–AOD (AI) slope and
CER3.7µm–AOD (AI) slope under all LWP bins over three anthropogenic
regions (EC, EU, and WE) and their adjacent oceans (ECO, EUO, and WEO).
Possible physical explanationsCollision and coalescence
In warm clouds, collision–coalescence is a much more effective process to
increase droplet size compared with diffusion of water vapor (Langmuir,
1948; Kogan, 1993; Beard and Ochs, 1993; Pruppacher and Klett, 1997). The
impact of collision–coalescence on CER is believed to be minimal when CER is
smaller than ∼14µm, while the coalescence rate
increases very rapidly when CER is greater than this value (Gerber, 1996;
Freud and Rosenfeld, 2012); i.e., the collision–coalescence becomes
dominant. To explore the possible contribution of collision–coalescence
processes to the correlation between CER and AOD/AI, the data are thus
stratified into two subsets with CER greater and less than 14 µm. As shown in Fig. 7, the dependence of CER on AOD/AI becomes
fairly weak with the slope close to zero over both land and ocean when the
collision–coalescence process is dominant (blue; CER >14µm), while CER–AOD/AI slopes are still positive over land and negative over
ocean when the collision–coalescence is negligible, which suggests that the
collision-coalescence is not likely the major cause for the opposite
CER–AOD/AI correlations over land and ocean. Figure 8 shows the probability
density functions of AOD/AI for the above two subsets over land and ocean. It is clearly demonstrated that the occurrence frequency of
AOD/AI presents a significant difference between the case of CER
>14µm and CER <14µm, with larger AOD for the
former and smaller AOD for the latter over land, while the reverse is true
over ocean. This result implies that collision–coalescence processes are
more likely to occur in the case of larger AOD over land and smaller AI over
ocean, which may further enhance the positive CER–AOD slope over the land
and the negative CER–AI slope over ocean, as demonstrated in Fig. 7. From the
above analyses, we assume that although the collision–coalescence is not the
dominant cause for the positive slope over land, the increased CER caused by
increased aerosol might further increase CER by initializing
collision–coalescence processes, generating a positive feedback.
Cloud effective radius (CER) varies with aerosol loading (AOD for
land, AI for ocean) under three liquid water path (LWP) categories
(20–60, 60–100, 100–140 g m-2). The figures from top to bottom represent different regions (EC,
EU, WE, ECO, EUO, and WEO). The blue (red) dots show the mean CER at each
category of AOD (AI) in the case of CER >14µm (<14µm), and the black circles indicate the results that do not
differentiate CER. The slopes on a log–log scale along with the total number
of samples (bracketed) for these three categories are also provided in each
panel.
Probability density functions of AOD (AI) for CER >14µm (blue) and CER <14µm (red) over three anthropogenic
regions (EC, EU, and WE) and their adjacent oceans (ECO, EUO, and WEO).
Entrainment mixing
In addition to the aforementioned collision–coalescence processes,
entrainment mixing is also believed to be an effective way to reduce cloud
droplet number concentration and increase CER, which of course depends on
the specific details of entrainment mixing mechanisms, as discussed in depth
by Kim et al. (2008). The possible impacts of the entrainment mixing process
on the estimated aerosol first indirect effect (AIE) have been reported by
previous studies. For instance, Kim et al. (2008) found that the AIE is
relatively weak in sub-adiabatic clouds compared with adiabatic clouds based
on ground-based remote sensing at the Southern Great Plains, and pointed out
that this may be due to interference from heterogeneous entrainment mixing
that changes the droplet number concentrations in a manner that attenuates
the AIE. Additionally, Shao and Liu (2006) suggested the AIE is about half
of that estimated by many previous studies, and attributed this difference
to evaporation resulting from entrainment mixing processes. By using a
mixed-layer model, Dal Gesso et al. (2014) found that the entrainment rate
in stratocumulus clouds tends to be larger for smaller lower
tropospheric stability (LTS). Although LTS is an indicator of temperature
inversion proposed to be relevant for marine stratocumulus clouds (Klein and
Hartmann, 1993), it has also been widely used to characterize atmospheric
stability over both land and ocean, and thereby indicating the vertical
mixing (Chen et al., 2014, 2016; Liu et al., 2017; Ma et al.,
2018). As already known, the degree of entrainment mixing is subject to not only the
intensity of turbulence in clouds but also the relative humidity of air
outside clouds (De Rooy et al., 2013). In this study, the former is
represented by lower tropospheric stability (LTS), and the latter is
described by relative humidity near cloud top (RHCT).
The slopes of CER versus AOD (AI) on a log–log scale under low,
medium, and high LTS and RHCT conditions, respectively, over (a) EC, (b) EU, (c) WE, (d) ECO, (e) EUO, and (f) WEO.
The range of LTS and RHCT for low,
medium, and high conditions.
To investigate the extent to which entrainment mixing processes affect the
slopes of CER versus AOD/AI, the data set is thus stratified according to LTS
and RHCT to differentiate environmental regimes. In order to satisfy
the constant LWP assumption for aerosol first indirect effect, a LWP bin range
from 60 to 100 g m-2 is employed, among which the LWP does not change
significantly with AOD/AI (as demonstrated in Fig. 1b) and statistical
samples are sufficient for regression analysis. The slopes of CER versus AOD
over land and the slopes of CER versus AI over ocean on a log–log scale under
low, medium, and high LTS and RHCT conditions are shown in Fig. 9, in
which nine categories are classified to make sure that each category has
roughly the same number of samples (Table 2). It is clearly shown that
slopes under low LTS (weak stable condition) are more positive than those
under medium (medium stable) and high LTS (stable condition) over both land
and ocean. Also, the slopes under low RHCT (dry condition) are overall
more positive than those under high RHCT (moist condition), although the
responses of slopes to RHCT are not as significant as those with LTS.
That is, under drier cloud top and stronger turbulence in clouds, positive
correlations more likely occur, which are associated with stronger
entrainment.
The distributions of sample number for (a, c) LTS and (b, d) RHCT
over land (a, b; EC, EU, and WE) and ocean (c, d; ECO, EUO, and WEO).
Vertical lines show the mean values. Texts with the corresponding color
represent the mean and standard deviation for each region.
Figure 10 shows the distributions of sample number for LTS and RHCT. It
is clearly demonstrated that there exists a systemic difference in both LTS
and RHCT between land and ocean; the frequencies of clouds under lower
LTS (Fig. 10a) and RHCT (Fig. 10b) are higher over land than over
ocean. The average LTS and RHCT over EC (10.3±2.3 and
67.8±14.8), EU (13.2±2.5 and 68.4±15.6), and WE
(12.5±3.2 and 65.1±16.0) are generally lower than that over ECO
(14.2±1.4 and 73.8±13.3), EUO (14.1±1.7 and 71.9±16.8), and WEO (17.9±3.1 and 75.8±23.5). Therefore, the
opposite correlations over major industrial regions and their adjacent
oceans discussed in Sect. 3.1 are probably associated with this systemic
difference in the indicators of entrainment mixing, i.e., LTS and RHCT.
As shown in Fig. 9, positive correlations are more likely to occur under the
condition of drier cloud top and stronger turbulence in clouds, which is
more prevalent over land, while negative correlations normally occur under
the condition of moister cloud top and weaker turbulence in clouds, which is
common over ocean. The opposite CER–AOD correlations over land and ocean
have also been generally found on a global scale (Grandey and Stier, 2010),
but this is not always the case for some regions (Nakajima et al., 2001;
Bréon et al., 2002), indicating that aerosol–cloud interactions remain
challenging due to the interference of other influencing factors, such as
different aerosol microphysical properties and coincidentally changing
meteorological conditions.
Summary
Using 14 years of co-located aerosol and cloud observations from MODIS
C6 L3 products, together with ERA-Interim reanalysis data, we systematically
investigated correlations between cloud effective radius (CER) and aerosol
optical depth (AOD) over three anthropogenic regions and correlations
between CER and aerosol index (AI) over their adjacent oceans, assessed the
effects of satellite retrieval biases on the aerosol–cloud correlations, and
then explored the underlying physical mechanisms. We also verified the
reliability of the CER–AOD/AI correlation derived from satellite data by
employing the CERES Edition-4 L3 TOA albedo product.
Our analysis indicated that cloud effective radius is overall positively
correlated with aerosol optical depth over land (positive slopes), but
negatively correlated with aerosol index over oceans (negative slopes).
Since CER is a function of both AOD/AI and LWP, the assumption of constant
LWP must be taken into consideration when estimating aerosol first indirect
effect. The results show that variations in LWP with AOD/AI in coarse LWP
bins (40 g m-2) can amplify the positive slope over land and the
negative slope over ocean. Therefore, we reanalyze the correlation between CER
and AOD/AI for narrow intervals of LWP (5 g m-2) in order to exclude
the interference of the covariation in LWP and AOD/AI, and still get
positive relationships over land and negative relationships over ocean.
The reliability of positive correlation between CER and AOD based on
satellite remote sensing is still controversial (Zhao et al., 2018) due to
the limitations of retrieval biases. Therefore, it is necessary to examine
whether the positive correlation over land is real before exploring possible
physical causes. We stratified data according to CF (with an interval of
0.04) and LWP (with an interval of 10 g m-2) to isolate the Twomey
effect. The results suggested that the changes in albedo at TOA
corresponding to aerosol-induced changes in CER also lend credence to the
reliability of positive correlation over land and negative correlation over ocean.
Although the positive correlations between CER and AOD over land are believed
to be a real relationship as discussed above, the magnitude of slope is
still subject to artificial correlations due to the retrieval biases in both
aerosol (cloud contamination and cloud adjacency effect) and cloud
(partially cloudy and 3-D-shaped clouds). To evaluate the contribution of
retrieval biases on artificial correlations, we compared the correlations
between CER and AOD/AI derived from PCL (partly cloudy) retrievals and normal
overcast retrievals, and took the impact of retrieval biases in aerosol
into account by stratifying data according to CF. It is suggested that CER
biases in PCL retrievals can induce an overestimation of the CER–AOD/AI
slope, and the retrieval biases of aerosol can amplify the overestimation of
slope caused by PCL retrievals. Additionally, our analysis shows that
retrieval biases for 3-D-shaped cloud also tend to result in a more positive
slope of CER versus AOD/AI than its physically correct value. Therefore, the
artificial positive correlation will be introduced if the following occurs:
(a) PCL pixels are erroneously identified as overcast pixels; (b) cloudy
pixels are erroneously identified as clear; and (c) 3-D-shaped clouds affect
both cloud retrievals in their own cloud pixels and aerosol retrievals in
surrounding cloud-free pixels.
We also explored potential physical mechanisms that can help to explain the
observed positive correlation between CER and AOD over land, including
collision–coalescence and entrainment mixing. It is suggested that the
collision–coalescence seems not to be the dominant cause for positive slope
of CER versus AOD over land, but there might be a positive feedback where
the increased CER caused by increased aerosol might further increase CER by
initializing collision–coalescence. Additionally, we stratified data
according to the lower tropospheric stability (LTS) and relative humidity
near cloud top (RHCT) to differentiate environmental regimes, and found
that the positive correlations more likely occur in the case of drier cloud top
and stronger turbulence in clouds, corresponding to stronger entrainment
mixing. Furthermore, there are systematic differences in LTS and
RHCT between land dominated by the positive correlation and ocean
dominated by the negative one; i.e., LTS and RHCT are lower over land
than over ocean. Therefore, it is inferred that entrainment mixing might be
a possible physical interpretation for such a positive CER–AOD slope.
It is acknowledged that although we have explored some uncertainties in
clarifying the correlation between aerosols and clouds, the complexity of
this issue and inherent limitations of polar-orbiting satellite measurements
make it difficult to reach a definitive conclusion regarding causal
relationships between CER and AOD/AI. Additional potential sources of the
uncertainties include the following.
Satellite-derived cloud-top CER might deviate from cloud-base or whole-cloud CER if the clouds are non-adiabatic, and/or coagulation of cloud
droplets occurs.
As a vertical integral quantity, AOD/AI is not always a good proxy for
cloud-base CCN concentrations in some cases (Stier, 2016).
AOD/AI might be loosely related to cloud-base CCN if affected by
deliquescence in high-relative-humidity environments, or if dust or other
insoluble aerosol has a substantial share of the total aerosol.
AOD/AI of clear-sky column retrievals might not be representative of the
cloud-base CCN in the neighboring cloudy pixels.
Unfortunately, it is quite difficult to evaluate these uncertainties in this
study due to the limitations of the data. However, our study still presents
sufficient evidence to warrant further investigations of the physical
mechanisms using detailed in situ field measurements of aerosol and cloud
properties, especially vertical profiles.
Data availability
The MODIS Aqua Level 3 Collection 6 product is available from https://ladsweb.modaps.eosdis.nasa.gov/archive/allData/6/MYD08_D3/ (last access:
9 March 2017; MODIS Science Team, 2014). The CERES SSF product is available from http://ceres.larc.nasa.gov/products.php?product=SSF1deg (last access: 5 October 2017; Doelling, 2016). The ECMWF ERA-Interim product is collected from the ECMWF data server http://www.ecmwf.int/en/research/climate-reanalysis/ (last access: 5 October 2017; Dee et al., 2011).
The supplement related to this article is available online at: https://doi.org/10.5194/acp-19-8879-2019-supplement.
Author contributions
HJ and XM designed the study and the statistical analysis. HJ processed the
satellite data and drafted the paper. XM, JQ, YY, and TQ validated and
debugged the results. All authors contributed to revising the paper.
Competing interests
The authors declare that they have no conflict of interest.
Acknowledgements
We are
grateful for the ease of access to MODIS and CERES, provided by NASA. We also
thank ECMWF for providing daily ERA-Interim reanalysis data in our work.
Financial support
This research has been supported by the National Key R&D Program of China (grant no. 2016YFA0600404), the National Natural Science Foundation of China (grant nos. 41675004 and 41590873), the Postgraduate Research & Practice Innovation Program of Jiangsu Province (grant no. KYCX18_1005), and the funding of Jiangsu innovation & entrepreneurship team.
Review statement
This paper was edited by Jianping Huang and reviewed by two anonymous referees.
ReferencesAlbrecht, B.: Aerosols, Cloud Microphysics, and Fractional Cloudiness,
Science, 245, 1227–1230, 10.1126/science.245.4923.1227, 1989.
Almeida, G. P., Brito, J., Morales, C. A., Andrade, M. F., and Artaxo, P.: Measured and modelled cloud condensation nuclei (CCN) concentration in São Paulo, Brazil: the importance of aerosol size-resolved chemical composition on CCN concentration prediction, Atmos. Chem. Phys., 14, 7559–7572, 10.5194/acp-14-7559-2014, 2014.
Beard, K. V. and Ochs, H. T.: Warm-Rain Initiation: An Overview of
Microphysical Mechanisms, J. Appl. Meteorol., 32, 608–625, 1993.
Bréon, F.-M., Tanré, D., and Generoso, S.: Aerosol effect on cloud
droplet size monitored from satellite, Science, 295, 834–838, 2002.Cahalan, R. F., Oreopoulos L., Wen G., Marshak A., Tsay S.-C., and DeFelice
T.: Cloud characterization and clear sky correction from Landsat 7, Remote
Sens. Environ., 78, 83–98, 10.1016/S0034-4257(01)00251-6, 2001.Chen, T. M., Guo, J. P., Li, Z. Q., Zhao, C., Liu, H., Cribb, M., Wang, F.,
and He, J.: A CloudSat perspective on the cloud climatology and its
association with aerosol perturbation in the vertical over East China, J.
Atmos. Sci., 73, 3599–3616, 10.1175/JAS-D-15-0309.1, 2016.Chen, Y.-C., Christensen, M. W., Stephens, G. L., and Seinfeld, J. H.:
Satellite-based estimate of global aerosol-cloud radiative forcing by marine
warm clouds, Nat. Geosci., 7, 643–646, 10.1038/ngeo2214,
2014.Christensen, M. W., Chen, Y.-C., and Stephens, G. L.: Aerosol indirect
effect dictated by liquid clouds, J. Geophys. Res., 121, 14636–14650,
10.1002/2016JD025245, 2016.
Coakley, J. A., Friedman, M. A., and Tahnk, W. R.: Retrieval of cloud
properties for partly cloudy imager pixels, J. Atmos. Ocean. Tech., 22,
3–17, 2005.Costantino, L. and Breon, F.-M.: Analysis of aerosol-cloud interaction from
multi-sensor satellite observations, Geophys. Res. Lett., 37, L11801,
10.1029/2009GL041828, 2010.Costantino, L. and Bréon, F.-M.: Aerosol indirect effect on warm clouds over South-East Atlantic, from co-located MODIS and CALIPSO observations, Atmos. Chem. Phys., 13, 69–88, 10.5194/acp-13-69-2013, 2013.
Dal Gesso, S., Siebesma, A. P., de Roode, R. S., and van Wessem, J. M.: A mixed-layer model perspective on stratocumulus
steadystates in a perturbed climate, Q. J. Roy. Meteor. Soc., 140, 2119–2131, 2014.Dee, D., Uppala, S., Simmons, A., Berrisford, P., Poli, P., Kobayashi, S.,
Andrae, U., Balmaseda, M., Balsamo, G., Bauer, P., Bechtold, P., Beljaars,
A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R.,
Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H.,
Hólm, E. V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi,
M., McNally, A. P., Monge-Sanz, B. M., Morcrette, J.-J., Park, B.-K.,
Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.-N., and Vitart, F.:
The ERA-Interim reanalysis: configuration and performance of the data
assimilation system, Q. J. Roy. Meteor. Soc., 137, 553–597, 10.1002/qj.828, 2011.de Rooy, W. C., Bechtold, P., Frohlich, K., Hohenegger, C., Jonker, H.,
Mironov, D., Siebesma, A. P., Teixeira, J., and Yano, J.-I.: Entrainment and
detrainment in cumulus convection: an overview, Q. J. Roy. Meteorol. Soc.,
139, 1–19, 10.1002/qj.1959, 2013.Doelling, D.: CERES Level 3 SSF1deg-Day Aqua-MODIS HDF file – Edition 4A, NASA Langley Atmospheric Science Data Center DAAC, 10.5067/aqua/ceres/ssf1degday_l3.004a, 2016.Doelling, D. R., Haney, C. O., Scarino, B. R., Gopalan, A., and Bhatt, R.:
Improvements to the Geostationary Visible Imager Ray-Matching Calibration
Algorithm for CERES Edition 4, J. Atmos. Ocean. Tech., 33, 2679–2698,
10.1175/jtech-d-16-0113.1, 2016a.Doelling, D. R., Sun, M., Nguyen, L. T., Nordeen, M. L., Haney, C. O.,
Keyes, D. F., and Mlynczak, P. E.: Advances in Geostationary-Derived Longwave
Fluxes for the CERES Synoptic (SYN1deg) Product, J. Atmos. Ocean. Tech., 33,
503–521, 10.1175/JTECH-D-15-0147.1, 2016b.Dusek, U., Frank, G. P., Hildebrandt, L., Curtius, J., Schneider, J.,
Walter, S., Chand, D., Drewnick, F., Hings, S., Jung, D., Borrmann, S., and
Andreae, M. O.: Size matters more than chemistry for cloud-nucleating
ability of aerosol particles, Science, 312, 1375–1378,
10.1126/science.1125261, 2006.Ervens, B., Feingold, G., and Kreidenweis, S. M.: Influence of water-soluble
organic carbon on cloud drop number concentration, J. Geophys. Res.,
110, 18211, 10.1029/2004JD005634, 2005.Fan, J., Yuan, T., Comstock, J. M., Ghan, S., Khain, A., Leung, L. R., Li,
Z., Martins, J. V., and Ovchinnikov, M.: Dominant role by vertical wind
shear in regulating aerosol effects on deep convective clouds, J. Geophys.
Res., 114, D22206, 10.1029/2009JD012352, 2009.
Feingold, G., Remer, L. A., Ramaprasad, J., and Kaufman, Y. J.: Analysis of
smoke impact on clouds in Brazilian biomass burning regions: An extension of
Twomey's approach, J. Geophys. Res., 106, 22907–22922, 2001.Feingold, G., Eberhard, W., Veron, D., and Previdi, M.: First measurements
of the Twomey indirect effect using ground-based remote sensors, Geophys.
Res. Lett., 30, 1287, 10.1029/2002GL016633, 2003.Freud, E. and Rosenfeld, D.: Linear relation between convective cloud drop number concentration and depth for rain initiation, J. Geophys. Res.-Atmos., 117, D02207, 10.1029/2011jd016457, 2012.Garrett, T. J., Zhao, C., Dong, X., Mace, G. G., and Hobbs, P. V.: Effects
of varying aerosol regimes on low-level Arctic stratus, Geophys. Res. Lett.,
31, L17105, 10.1029/2004GL019928, 2004.
Gerber, H.: Microphysics of marine stratocumulus clouds with two drizzle
modes, J. Atmos. Sci., 53, 1649–1662, 1996.Grandey, B. S. and Stier, P.: A critical look at spatial scale choices in satellite-based aerosol indirect effect studies, Atmos. Chem. Phys., 10, 11459–11470, 10.5194/acp-10-11459-2010, 2010.Grosvenor, D. P., Sourdeval, O., Zuidema, P., Ackerman, A., Alexandrov, M.
D., Bennartz, R., Boers, R., Cairns, B., Chiu, J. C., Christensen, M.,
Deneke, H., Diamond, M., Feingold, G., Fridlind, A., Hünerbein, A.,
Knist, C., Kollias, P., Marshak, A., McCoy, D., Merk, D., Painemal, D.,
Rausch, J., Rosenfeld, D., Russchenberg, H., Seifert, P., Sinclair, K.,
Stier, P., van Diedenhoven, B., Wendisch, M., Werner, F., Wood, R., Zhang,
Z., and Quaas, J.: Remote Sensing of Droplet Number Concentration in Warm
Clouds: A review of the current state of knowledge and perspectives, Rev.
Geophys., 56, 409–453,, 10.1029/2017RG000593, 2018.Gryspeerdt, E. and Stier, P.: Regime-based analysis of aerosol-cloud
interactions, Geophys. Res. Lett., 39, L21802, 10.1029/2012GL053221, 2012.Gryspeerdt, E., Stier, P., and Grandey, B.: Cloud fraction mediates the
aerosol optical depth – cloud top height relationship, Geophys. Rev. Lett.,
41, 3622–3627, 10.1002/2014GL059524, 2014.Gryspeerdt, E., Quaas, J., and Bellouin, N.: Constraining the aerosol
influence on cloud fraction, J. Geophys. Res., 121, 3566–3583,
10.1002/2015JD023744, 2016.Gryspeerdt, E., Quaas, J., Ferrachat, S., Gettelman, A., Ghan, S., Lohmann,
U., Morrison, H., Neubauer, D., Partridge, D. G., Stier, P., Takemura, T.,
Wang, H., Wang, M., and Zhang, K.: Constraining the instantaneous aerosol
influence on cloud albedo, P. Natl. Acad. Sci. USA, 114, 4899–4904,
10.1073/pnas.1617765114, 2017.Guo, J., Liu, H., Li, Z., Rosenfeld, D., Jiang, M., Xu, W., Jiang, J. H., He, J., Chen, D., Min, M., and Zhai, P.: Aerosol-induced changes in the vertical structure of precipitation: a perspective of TRMM precipitation radar, Atmos. Chem. Phys., 18, 13329–13343, 10.5194/acp-18-13329-2018, 2018.
Han, Q. Y., Rossow, W. B., and Lacis, A. A.: Near-global survey of effective
droplet radii in liquid water clouds using ISCCP data, J. Climate, 7,
465–497, 1994.
Huang, J., Guo, J., Wang, F., Liu, Z., Jeong, M. J., Yu, H., and Zhang, Z.:
CALIPSO inferred most probable heights of global dust and smoke layers, J.
Geophys. Res.-Atmos., 120, 5085–5100, 2015.IPCC: Climate Change 2013: The Physical Science Basis. Contribution of
Working Group I to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change, edited by: Stocker, T. F., Qin, D., Plattner,
G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V.,
and Midgley, P. M., Cambridge University Press, Cambridge, UK
and New York, NY, USA, 1535 pp., 10.1017/CBO9781107415324, 2013.Jin, M. and Shepherd, J. M.: Aerosol relationships to warm season clouds and
rainfall at monthly scales over east China: Urban land versus ocean, J.
Geophys. Res., 113, D24S90, 10.1029/2008JD010276, 2008.Kaufman, Y. J., Koren, I., Remer, L. A., Rosenfeld, D., and Rudich, Y.: The
effect of smoke, dust, and pollution aerosol on shallow cloud development
over the Atlantic Ocean, P. Natl. Acad. Sci. USA, 102, 11207–11212, 10.1073/pnas.0505191102, 2005.Kim, B.-G., Schwartz, S. E., Miller, M. A., and Min, Q.: Effective radius of
cloud droplets by ground-based remote sensing: Relationship to aerosol, J.
Geophys. Res., 108, 4740, 10.1029/2003JD003721, 2003.Kim, B.-G., Miller, M. A., Schwartz, S. E., Liu, Y., and Min, Q.: The role
of adiabaticity in the aerosol first indirect effect, J. Geophys. Res., 113,
D05210, 10.1029/2007JD008961, 2008.
Klein, S. A. and Hartmann, D. L.: The seasonal cycle of low stratiform
clouds, J. Climate, 6, 1588–1606, 1993.Kleinman, L. I., Daum, P. H., Lee, Y.-N., Lewis, E. R., Sedlacek III, A. J., Senum, G. I., Springston, S. R., Wang, J., Hubbe, J., Jayne, J., Min, Q., Yum, S. S., and Allen, G.: Aerosol concentration and size distribution measured below, in, and above cloud from the DOE G-1 during VOCALS-REx, Atmos. Chem. Phys., 12, 207–223, 10.5194/acp-12-207-2012, 2012.Koren, I., Kaufman, Y. J., Rosenfeld, D., Remer, L. A., and Rudich, Y.:
Aerosol invigoration and restructuring of Atlantic convective clouds,
Geophys. Res. Lett., 32, L14828, 10.1029/2005GL023187, 2005.Koren, I., Feingold, G., and Remer, L. A.: The invigoration of deep convective clouds over the Atlantic: aerosol effect, meteorology or retrieval artifact?, Atmos. Chem. Phys., 10, 8855–8872, 10.5194/acp-10-8855-2010, 2010.
Kogan, Y. L.: Drop Size Separation in Numerically Simulated Convective
Clouds and Its Effect on Warm Rain Formation, J. Atmos. Sci., 50,
1238–1253, 1993.Lance, S., Nenes, A., and Rissman, T. A.: Chemical and dynamical effects on
cloud droplet number: Implications for estimates of the aerosol indirect
effect, J. Geophys. Res., 109, D22208, 10.1029/2004JD004596, 2004.
Langmuir, I.: The production of rain by a chain reaction in cumulus clouds
at temperatures above freezing, J. Atmos. Sci., 5, 175–192, 1948Levy, R. C., Remer, L. A., Kleidman, R. G., Mattoo, S., Ichoku, C., Kahn, R., and Eck, T. F.: Global evaluation of the Collection 5 MODIS dark-target aerosol products over land, Atmos. Chem. Phys., 10, 10399–10420, 10.5194/acp-10-10399-2010, 2010.Levy, R. C., Mattoo, S., Munchak, L. A., Remer, L. A., Sayer, A. M., Patadia, F., and Hsu, N. C.: The Collection 6 MODIS aerosol products over land and ocean, Atmos. Meas. Tech., 6, 2989–3034, 10.5194/amt-6-2989-2013, 2013.Li, Z., Lau, W. M., Ramanathan, V., Wu, G., Ding, Y., Manoj, M. G., Liu, J.,
Qian, Y., Li, J., Zhou, T., Fan, J., Rosenfeld, D., Ming, Y., Wang, Y.,
Huang, J., Wang, B., Xu, X., Lee, S.-S., Cribb, M., Zhang, F., Yang, X.,
Zhao, C., Takemura, T., Wang, K., Xia, X., Yin, Y., Zhang, H., Guo, J.,
Zhai, P. M., Sugimoto, N., Babu, S. S., and Brasseur, G. P.: Aerosol and
monsoon climate interactions over Asia, Rev. Geophys., 54, 866–929,
10.1002/2015RG000500, 2016.Lihavainen, H., Kerminen, V.-M., and Remer, L. A.: Aerosol-cloud interaction determined by both in situ and satellite data over a northern high-latitude site, Atmos. Chem. Phys., 10, 10987–10995, 10.5194/acp-10-10987-2010, 2010.Liu, Y., de Leeuw, G., Kerminen, V.-M., Zhang, J., Zhou, P., Nie, W., Qi, X., Hong, J., Wang, Y., Ding, A., Guo, H., Krüger, O., Kulmala, M., and Petäjä, T.: Analysis of aerosol effects on warm clouds over the Yangtze River Delta from multi-sensor satellite observations, Atmos. Chem. Phys., 17, 5623–5641, 10.5194/acp-17-5623-2017, 2017.
Loeb, N. G., Kato, S., Loukachine, K., and Manalo-Smith, N.: Angular
distribution models for top-of-atmosphere radiative flux estimation from the
Clouds and Earth's Radiant Energy System instrument on the Terra satellite.
Part 1 methodogy, J. Atmos. Ocean. Tech., 22, 338–351, 2005.
Lohmann, U. and Lesins, G.: Stronger constraints on the anthropogenic
indirect aerosol effect, Science, 298, 1012–1016, 2002.Lu, C., Liu Y., Niu S., and Vogelmann A. M.: Observed impacts of vertical
velocity on cloud microphysics and implications for aerosol indirect
effects, Geophys. Res. Lett., 39, L21808, 10.1029/2012GL053599, 2012.
Ma, X., Jia, H., Yu, F., and Quaas, J.: Opposite aerosol index-cloud droplet
effective radius correlations over major industrial regions and their
adjacent oceans, Geophys. Res. Lett., 45, 5771–5778, 2018.
Manoj, M. G., Devara, P. C. S., Joseph, S., and Sahai, A. K.: Aerosol
indirect effect during the aberrant Indian Summer Monsoon breaks of 2009,
Atmos. Environ., 60, 153–163, 2012.Martins, J. V., Tanre, D., Remer, L. A., Kaufman, Y. J., Mattoo, S., and Levy,
R.: MODIS cloud screening for remote sensing of aerosol over oceans using
spatial variability, Geophys. Res. Lett., 29, 8009, 10.1029/2001GL013252,
2002.Matheson, M. A., Coakley Jr., J. A., and Tahnk, W. R.: Multiyear advanced
very high resolution radiometer observations of summertime stratocumulus
collocated with aerosols in the northeastern Atlantic, J. Geophys. Res.,
111, D15206, 10.1029/2005JD006890, 2006.McFiggans, G., Artaxo, P., Baltensperger, U., Coe, H., Facchini, M. C., Feingold, G., Fuzzi, S., Gysel, M., Laaksonen, A., Lohmann, U., Mentel, T. F., Murphy, D. M., O'Dowd, C. D., Snider, J. R., and Weingartner, E.: The effect of physical and chemical aerosol properties on warm cloud droplet activation, Atmos. Chem. Phys., 6, 2593–2649, 10.5194/acp-6-2593-2006, 2006.Mielonen, T., Levy, R. C., Aaltonen, V., Komppula, M., de Leeuw, G., Huttunen, J., Lihavainen, H., Kolmonen, P., Lehtinen, K. E. J., and Arola, A.: Evaluating the assumptions of surface reflectance and aerosol type selection within the MODIS aerosol retrieval over land: the problem of dust type selection, Atmos. Meas. Tech., 4, 201–214, 10.5194/amt-4-201-2011, 2011.MODIS Science Team: MYD08_D3 MODIS/Aqua Aerosol Cloud Water Vapor Ozone Daily L3 Global 1Deg CMG, Level 1 and Atmosphere Archive and Distribution System (LAADS), 10.5067/modis/myd08_d3.006, 2014.
Nakajima, T. and King, M. D.: Determination of optical thickness and
effective particle radius of clouds from reflected solar radiation
measurements. Part I: Theory, J. Atmos. Sci., 47, 1878–1893, 1990.
Nakajima, T., Higurashi, A., Kawamoto, K., and Penner, J. E.: A possible
correlation between satellite-derived cloud and aerosol microphysical
parameters, Geophys. Res. Lett., 28, 1171–1174, 2001.Nenes, A., Charlson, R. J., Facchini, M. C., Kulmala, M., Laaksonen, A., and
Seinfeld, J. H.: Can chemical effects on cloud droplet number rival the
first indirect effect? Geophys. Res. Lett., 29, 1848, 10.1029/2002GL015295, 2002.
Panicker, A. S., Pandithurai, G., and Dipu, S.: Aerosol indirect effect
during successive contrasting monsoon seasons over Indian subcontinent using
MODIS data, Atmos. Environ., 44, 1937–1943, 2010.Pawlowska, H. and Brenguier, J. L.: Microphysical properties of
stratocumulus clouds during ACE-2, Tellus B, 52, 868–887,10.1034/j.1600-0889.2000.00076.x, 2000.
Peng, J., Li, Z., Zhang, H., Liu, J., and Cribb, M.: Systematic changes in cloud radiative forcing with aerosol loading for deep clouds in the tropics, J. Atmos. Sci., 73, 231–249, 2016.
Platnick, S.: Vertical photon transport in cloud remote sensing problems, J.
Geophys. Res., 105, 22919–22935, 2000.Platnick, S., Meyer, K. G., King, M. D., Wind, G., Amarasinghe, N.,
Marchant, B., Arnold, G. T., Zhang, Z., Hubanks, P. A., Holz, R. E., Yang,
P., Ridgway, W. L., and Riedi, J.: The MODIS cloud optical and microphysical
products: Collection 6 updates and examples from Terra and Aqua, IEEE T.
Geosci. Remote, 55, 502–525, 10.1109/TGRS.2016.2610522, 2017.
Pruppacher, H. R. and Klett, J. D.: Microphysics of Clouds and
Precipitation, 2nd ed., Kluwer Academic, London, UK, 954 pp., 1997.Qiu, Y., Zhao, Z., Guo, J., and Li, J.: 8-Year ground-based observational
analysis about the seasonal variation of the aerosol-cloud droplet effective
radius relationship at SGP site, Atmos. Environ., 164, 139–146, 10.1016/j.atmosenv.2017.06.002, 2017.Quaas, J., Boucher, O., Bellouin, N., and Kinne, S.: Satellite-based
estimate of the direct and indirect aerosol climate forcing, J. Geophys.
Res., 113, D05204, 10.1029/2007JD008962, 2008.Quaas, J., Ming, Y., Menon, S., Takemura, T., Wang, M., Penner, J. E., Gettelman, A., Lohmann, U., Bellouin, N., Boucher, O., Sayer, A. M., Thomas, G. E., McComiskey, A., Feingold, G., Hoose, C., Kristjánsson, J. E., Liu, X., Balkanski, Y., Donner, L. J., Ginoux, P. A., Stier, P., Grandey, B., Feichter, J., Sednev, I., Bauer, S. E., Koch, D., Grainger, R. G., Kirkevåg, A., Iversen, T., Seland, Ø., Easter, R., Ghan, S. J., Rasch, P. J., Morrison, H., Lamarque, J.-F., Iacono, M. J., Kinne, S., and Schulz, M.: Aerosol indirect effects – general circulation model intercomparison and evaluation with satellite data, Atmos. Chem. Phys., 9, 8697–8717, 10.5194/acp-9-8697-2009, 2009.Quaas, J., Stevens, B., Stier, P., and Lohmann, U.: Interpreting the cloud cover – aerosol optical depth relationship found in satellite data using a general circulation model, Atmos. Chem. Phys., 10, 6129–6135, 10.5194/acp-10-6129-2010, 2010.
Remer, L. A., Kaufman, Y. J., Tanre, D., Mattoo, S., Chu, D. A., Martins, J.
V., Li, R. R., Ichoku, C., Levy, R. C., Kleidman, R. G., Eck, T. F.,
Vermote, E., and Holben, B. N.: The MODIS aerosol algorithm, products, and
validation, J. Atmos. Sci., 62, 947–973, 2005.Roberts, G. C., Ramana, M. V., Corrigan, C., Kim, D., and Ramanathan, V.:
Simultaneous observations of aerosol-cloudalbedo interactions with three
stacked unmanned aerial vehicles, P. Natl. Acad. Sci., 105, 7370–7375,
10.1073/pnas.0710308105, 2008.
Robock, A.: The seasonal cycle of snow cover, sea ice and surface albedo,
Mon. Weather Rev., 108, 267–285, 1980.Saponaro, G., Kolmonen, P., Sogacheva, L., Rodriguez, E., Virtanen, T., and de Leeuw, G.: Estimates of the aerosol indirect effect over the Baltic Sea region derived from 12 years of MODIS observations, Atmos. Chem. Phys., 17, 3133–3143, 10.5194/acp-17-3133-2017, 2017.Sayer, A. M., Hsu, N. C., Bettenhausen, C., and Jeong, M.-J.: Validation and
uncertainty estimates for MODIS Collection 6 “Deep Blue” aerosol data, J.
Geophys. Res., 118, 7864–7872, 10.1002/jgrd.50600, 2013.Sekiguchi, M., Nakajima, T., Suzuki, K., Kawamoto, K., Higurashi, A.,
Rosenfeld, D., Sano, I., and Mukai, S.: A study of the direct and indirect
effects of aerosols using global satellite data sets of aerosol and cloud
parameters, J. Geophys. Res., 108, 4699, 10.1029/2002JD003359,
2003.Shao, H. and Liu, G.: Influence of mixing on evaluation of the aerosol
first indirect effect, Geophys. Res. Lett., 33, L14809,
10.1029/2006GL026021, 2006.Stier, P.: Limitations of passive remote sensing to constrain global cloud condensation nuclei, Atmos. Chem. Phys., 16, 6595–6607, 10.5194/acp-16-6595-2016, 2016.Su, W., Loeb, N. G., Xu, K. M., Schuster, G. L., and Eitzen, Z. A.: An
estimate of aerosol indirect effect from satellite measurements with
concurrent meteorological analysis, J. Geophys. Res., 115, D18219,
10.1029/2010JD013948, 2010.Tang, J., Wang, P., Mickley, L. J., Xia, X., Liao, H., Yue, X., Sun, L., and
Xia, J.: Positive relationship between liquid cloud droplet effective radius
and aerosol optical depth over Eastern China from satellite data, Atmos.
Environ., 84, 244–253, 10.1016/j.atmosenv.2013.08.024, 2014.Tripathi, S. N., Dey, S., Chandel, A., Srivastava, S., Singh, R. P., and Holben, B. N.: Comparison of MODIS and AERONET derived aerosol optical depth over the Ganga Basin, India, Ann. Geophys., 23, 1093–1101, 10.5194/angeo-23-1093-2005, 2005.
Twomey, S.: The influence of pollution on the shortwave albedo of clouds, J.
Atmos. Sci., 34, 1149–1152, 1977.Vant-Hull, B., Marshak, A., Remer, L. A., and Li, Z. Q.: The effects of
scattering angle and cumulus cloud geometry on satellite retrievals of cloud
droplet effective radius, IEEE T. Geosci. Remote, 45, 1039–1045,
10.1109/TGRS.2006.890416, 2007.Várnai, T. and Marshak, A.: MODIS observations of enhanced clear-sky
reflectance near clouds, Geophys. Res. Lett., 36, L06807,
10.1029/2008GL037089, 2009.Wall, C., Zipser, E., and Liu, C. T.: An investigation of the aerosol
indirect effect on convective intensity using satellite observations, J.
Atmos. Sci., 71, 430–447, 10.1175/jas-d-13-0158.1, 2014.Wang, F., Guo, J., Wu, Y., Zhang, X., Deng, M., Li, X., Zhang, J., and Zhao,
J.: Satellite observed aerosol-induced variability in warm cloud properties
under different meteorological conditions over eastern China, Atmos.
Environ., 84, 122–132, 10.1016/j.atmosenv.2013.11.018, 2014.Wang, F., Guo, J., Zhang, J., Huang, J., Min, M., Chen, T., Liu, H., Deng,
M., and Li, X.: Multi-sensor quantification of aerosolinduced variability in
warm clouds over eastern China, Atmos. Environ., 113, 1–9,
10.1016/j.atmosenv.2015.04.063, 2015.Wang, J., Lee, Y.-N., Daum, P. H., Jayne, J., and Alexander, M. L.: Effects of aerosol organics on cloud condensation nucleus (CCN) concentration and first indirect aerosol effect, Atmos. Chem. Phys., 8, 6325–6339, 10.5194/acp-8-6325-2008, 2008.Wang, J., Cubison, M. J., Aiken, A. C., Jimenez, J. L., and Collins, D. R.: The importance of aerosol mixing state and size-resolved composition on CCN concentration and the variation of the importance with atmospheric aging of aerosols, Atmos. Chem. Phys., 10, 7267–7283, 10.5194/acp-10-7267-2010, 2010.Wang, K., Liu, J., Zhou, X., Sparrow, M., Ma, M., Sun, Z., and Jiang, W.:
Validation of the MODIS global land surface albedo product using ground
measurements in a semidesert region on the Tibetan Plateau, J. Geophys.
Res., 109, D05107, 10.1029/2003JD004229, 2004.
Wang, Z., Vane, D., Stephens, G., and Reinke, D.: CloudSat Project: Level 2
Combined Radar and Lidar Cloud Scenario Classification Product Process
Description and Interface Control Document, California Institute of
Technology, California, USA, 61 pp., 2013.
Wen, G., Marshak, A., and Cahalan, R. F.: Impact of 3-D clouds on clear sky
reflectance and aerosol retrieval in a biomass burning region of Brazil,
IEEE Geosci. Remote, 3, 169–172, 2006.Wen, G., Marshak, A., Cahalan, R., Remer, L., and Kleidman, R.: 3-D
aerosol-cloud radiative interaction observed in collocated MODIS and ASTER
images of cumulus cloud fields, J. Geophys. Res.-Atmos., 112, D13204,
10.1029/2006JD008267, 2007.Werner, F., Ditas, F., Siebert, H., Simmel, M., Wehner, B., Pilewskie, P.,
Schmeissner, T., Shaw, R. A., Hartmann, S., Wex, H., Roberts, G. C., and
Wendisch, M.: Twomey effect observed from collocated microphysical and
remote sensing measurements over shallow cumulus, J. Geophys. Res.-Atmos.,
119, 1534–1545, 10.1002/2013JD020131, 2014.
Wielicki, B. A., Barkstrom, B. R., Harrison, E. F., Lee, R. B. I., Smith, G.
L., and Cooper, J. E.: Clouds and the Earth's radiant energy system (CERES):
An Earth observing system experiment, B. Am. Meteorol. Soc., 77,
853–868, 1996.Wilcox, E. M., Roberts, G., and Ramanathan, V.: Influence of aerosols on the
shortwave cloud radiative forcing from north Pacific Oceanic Clouds: Results
from the Cloud Indirect Forcing Experiment (CIFEX), Geophys. Res. Lett., 33,
L21804, 10.1029/2006GL027150, 2006.Yuan, T., Li, Z., Zhang, R., and Fan, J.: Increase of cloud droplet size
with aerosol optical depth: An observation and modeling study, J. Geophys.
Res., 113, D04201, 10.1029/2007JD008632, 2008.
Zhang, J., Reid, J., and Holben, B.: An analysis of potential cloud
artifacts in MODIS over ocean aerosol optical thickness products, Geophys.
Res. Lett., 32, L15803, 10.1029/2005GL023254, 2005.Zhang, Q., Quan, J., Tie, X., Huang, M., and Ma, X.: Impact of aerosol
particles on cloud formation: Aircraft measurements in China, Atmos.
Environ., 45, 665–672, 10.1016/j.atmosenv.2010.10.025, 2011.Zhang, Z. and Platnick S.: An assessment of differences between cloud
effective particle radius retrievals for marine water clouds from three
MODIS spectral bands, J. Geophys. Res., 116, D20215,
10.1029/2011JD016216, 2011.Zhang, Z., Ackerman, A. S., Feingold, G., Platnick, S., Pincus, R., and Xue,
H.: Effects of cloud horizontal inhomogeneity and drizzle on remote sensing
of cloud droplet effective radius: Case studies based on large-eddy
simulations, J. Geophys. Res., 117, D19208, 10.1029/2012JD017655, 2012.Zhao, C. F., Qiu, Y. M., Dong, X. B., Wang, Z. E., Peng, Y. R., Li, B. D., Wu, Z. H., and Wang, Y.: Negative Aerosol-Cloud re Relationship From Aircraft Observations Over Hebei, China, Earth Space Sci., 5, 19–29, 10.1002/2017ea000346, 2018.Zuidema, P., Redemann, J., Haywood, J., Wood, R., Piketh, S., Hipondoka, M.,
and Formenti, P.: Smoke and Clouds above the Southeast Atlantic: Upcoming
Field Campaigns Probe Absorbing Aerosol's Impact on Climate, B. Am.
Meteorol. Soc., 97, 1131–1135, 10.1175/BAMS-D-15-00082.1, 2016.